Election and Adultery Special

Episode Summary

Executive Summary: BBC Radio 4’s More or Less hosts a live, unscripted analysis of the 2015 UK election, where experts debate whether pollsters failed, why the exit poll was more accurate, and how media and politics misuse numbers. The show then pivots to a dubious adultery ranking based on dating-site data, before ending with a discussion of election fact-checking and the difficulty of translating political claims into meaningful figures.

Main Topics: Election polls vs. election result (Priority: 5/5): The panel examines why pre-election polls suggested a hung parliament while the Conservatives won decisively, with debate over whether pollsters were wrong, cautious, or constrained by methodology. Exit poll accuracy and polling methods (Priority: 5/5): Participants explain why the BBC/ITV/Sky exit poll performed better than opinion polls: it sampled actual voters after they voted and measured change rather than intention. Systemic polling bias and ‘herding’ (Priority: 4/5): Experts argue that many pollsters overstated Labour and understated Conservatives, possibly due to consistent bias and pollsters aligning forecasts with one another and with media expectations. Statistics of adultery and misleading media numbers (Priority: 4/5): A segment on Beeston being labelled the UK’s most adulterous town exposes how self-selecting dating-site data can be misrepresented as social research. Live fact-checking during the campaign (Priority: 4/5): Guests from Full Fact, the IFS, and the Mirror discuss the growing demand for real-time fact-checking and the challenges of clarifying vague or conflicting political claims. Politics, rhetoric, and numerical ambiguity (Priority: 3/5): The show highlights how politicians can use numbers rhetorically—such as different claims about household gains or losses—without technically lying, but still confusing the public.

Key Arguments: Polls are not pure predictions; they measure stated intention at a specific moment, and small shifts can change results dramatically in the UK electoral system. The 2015 polling failure was not random noise alone: several pollsters systematically overstated Labour and understated Conservatives, suggesting a directional bias. Margin of error is an insufficient defense when multiple polls point in the same wrong direction across an election campaign. Exit polls are more reliable because they survey actual voters immediately after voting, reducing uncertainty about turnout and choice. Media and pollsters may reinforce each other through ‘herding,’ with pollsters reluctant to diverge from the consensus and media overstating small movements. The adultery ranking based on a dating site is not a valid measure of real affairs because it reflects site registration, not actual behavior. Fact-checking has become mainstream in politics and media, but it works best when it clarifies what claims actually mean rather than just labeling them true or false. Political parties often use vague language or incompatible definitions to make numbers sound contradictory when they are actually talking about different things.

Data Points: Conservative vote share (Trumptonshire mock result): 23,252 votes - Election-style satirical result read out in the live programme Labour vote share (Trumptonshire mock result): 14,246 votes - Mock constituency result used as a humorous opening UKIP vote share (Trumptonshire mock result): 7,133 votes - Mock election result Liberal Democrat vote share (Trumptonshire mock result): 1,827 votes - Mock election result Green vote share (Trumptonshire mock result): 1,370 votes - Mock election result UKIP national share: 13% - Ben Page notes his sock-eating dare was missed because UKIP did not reach 18% UKIP dare threshold: 18% - Ben Page’s pre-election promise to eat his socks if UKIP reached this share Conservative poll estimate by Ipsos Mori on Wednesday: 36% - Ben Page says their final poll was close to the Conservative result Conservative actual vote share: 37.8% - Ben Page says Conservatives were within margin of error of Ipsos Mori poll Polling companies reviewed by BBC: 9 companies - David Cowling says he examined the last poll from nine firms Polls overstating Labour: 9 out of 9 pollsters - Cowling says all nine overstated Labour to varying degrees Polls understating Conservatives: 8 out of 9 pollsters - Cowling says eight pollsters understated Conservatives Typical polling margin of error: about plus or minus 3% - David Spiegelhalter describes a 1,000-sample poll Exit poll sampling points: about 140 contact points - Cowling explains the design of the BBC exit poll Polling districts sampled in exit poll: 39,000 polling districts - Exit poll methodology description Beeston adultery claim: 2.5% of population - Media claim about residents allegedly having an affair Beeston reported affairs: 941 affairs - Claim cited in the adultery segment Women registered on the dating site: 80,000 - Spedra? Actually the discussion says the site has 80,000 women registered in the country IFS estimate for Labour current budget balance implications: £25–30 billion a year - Paul Johnson explains the likely investment spending assumption Labour’s implied deficit reduction target: from £90 billion to about £30 billion - IFS interpretation of Labour’s vague fiscal plan Household gain claim by Osborne: £900 a year better off - Example of conflicting campaign numbers Household loss claim by Miliband: £1,600 a year worse off - Example showing different definitions behind competing claims

Pivotal Quotes: "Polls don't predict. Pollsters and the people running them do." — Ben Page: He argues that polls measure current intentions rather than forecast outcomes "I don't think margin of error is something you can get away with in this case." — Sir David Spiegelhalter: He rejects the idea that polling errors were excused by statistical uncertainty "It wasn't that the politicians got their facts wrong. It was that their facts were just profoundly unhelpful or misleading." — Tim Harford: Summarizing the problem with campaign numerics and rhetoric

Implications: Listeners are urged to treat polls, viral stats, and political claims cautiously: methods matter, definitions matter, and numbers can mislead even when technically accurate. Fact-checking and transparent methodology are essential for public trust.